Head and Shoulders Pattern: Tops, Inverse Bottoms and Confirmation

Where this page's information comes from

• StockCharts' ChartSchool, its guides to the head and shoulders top and bottom, for the conventions on anatomy, volume, retests and targets. Those points are attributed in the text.
• Thomas Bulkowski's glossary on ThePatternSite, for his definitions of the neckline, pullbacks, throwbacks, the measure rule and busted patterns.
• TradingLoot's own explanations and diagrams. They describe how the pattern is commonly read, not rules that guarantee an outcome.

This page gives no success rates or performance statistics. For how Bulkowski measured patterns, see our summary of his encyclopedia.

Anatomy of a head and shoulders top

It starts with the trend. ChartSchool is explicit that without a prior uptrend to reverse, there can't be a head and shoulders top. A similar shape in the middle of a range is just a shape.

ChartSchool describes the parts in sequence. The left shoulder is a peak within the uptrend, and its pullback usually stays above the trend line. The head is a higher peak, and its decline generally breaks below the uptrend line, which puts the trend in doubt. The right shoulder is a lower high, usually roughly in line with the left shoulder. Each step is a small piece of evidence that buyers are losing their grip, but none of them completes the pattern on its own.

Our summary of Bulkowski's encyclopedia notes that the book gives complex versions, with extra shoulders or heads, chapters of their own. The reading on this page is the simple form.

The neckline and confirmation

The neckline joins the two lows between the peaks. Bulkowski describes it as the line through the pattern's "armpits", extended forward so that a signal comes when price crosses it. It can slope up, slope down or be flat. ChartSchool considers a downward-sloping neckline more bearish than an upward one; treat that as its interpretation, not a rule.

The neckline is also where the pattern is confirmed. Until price breaks it convincingly, the uptrend hasn't been reversed, and you have a possible head and shoulders, not a completed one. A three-peak shape may never break its neckline at all. The same principle applies to a double top: the low between its two peaks defines the support level whose break is used for confirmation. What counts as "convincing" varies between sources; the hub covers common filters.

Volume

ChartSchool reads volume as a series of warning signs in a top. Lighter volume as the head makes its new high is one; heavier volume on the decline from the head, then lighter volume as the right shoulder rallies, is another. It then looks for volume to expand on the break of the neckline as confirmation.

These are tendencies that a source describes, not conditions every pattern meets. Volume is supporting evidence to weigh alongside the neckline break, and it carries more weight in the inverse pattern than in the top, as the section on the inverse explains.

Retests of the neckline

Once the neckline breaks, ChartSchool notes that the old support level commonly turns into resistance, and that price sometimes, but certainly not always, comes back to it. Bulkowski's term for that return after a downward breakout is a pullback, which he defines as a return to the breakout price within 30 days.

Two cautions follow. A retest isn't guaranteed, so waiting for one can mean missing the move entirely. And a return to the neckline that keeps going back up through it isn't a retest any more: it's the start of a failed pattern.

The measured-move convention

The usual convention, as ChartSchool describes it, measures the distance from the head down to the neckline and subtracts it from the point where the neckline broke. That is marker 7 in the diagram. Bulkowski calls this kind of calculation a measure rule, describing a minimum price move after the breakout.

ChartSchool presents the result as a rough guide and suggests weighing it against other factors, such as earlier support levels, Fibonacci retracements and moving averages. It's a reference point, not a prediction of where price will go or how quickly. For why TradingLoot doesn't publish hit rates for targets, see probabilities and limits.

Inverse head and shoulders

The inverse head and shoulders, also called a head and shoulders bottom, is the same pattern upside down: after a downtrend, a low, a lower low and a higher low, with the neckline drawn through the two highs between them. ChartSchool stresses that it needs a prior downtrend, for the same reason the top needs an uptrend. It is complete only when price breaks above the neckline convincingly.

The main difference is volume. ChartSchool says an increase in volume on the neckline breakout is "welcomed" for a top but "absolutely required for a bottom", and it treats volume in the second half of the pattern, from the head onward, as more important than in the first. Without that expansion, it regards the breakout as suspect.

Retests work the same way in reverse: broken resistance often becomes support, and ChartSchool notes that price often returns to the broken neckline. Bulkowski calls a return after an upward breakout a throwback.

Similar patterns

Tops with three peaks are also covered in the triple top guide, which compares the two shapes:

• Triple tops and bottoms compared with head and shoulders and double tops

Failed patterns and false signals

A head and shoulders can go wrong at two stages. Before confirmation, the shape simply doesn't complete: price holds the neckline and the uptrend carries on. That isn't a failure of the pattern, because there never was a confirmed pattern. After confirmation, price can break the neckline, find little follow-through and come back above it. That is a failed breakout.

A failed breakout is a broad idea. Bulkowski's busted pattern is a narrower, specific definition: price moves no more than 10% after the breakout, then reverses and crosses the other side of the pattern, starting a trend in the new direction. So not every failed breakout is a busted pattern in his sense. For why failures still carry information, see our summary of Bulkowski's research on failure.

False signals also come from reading too early: calling a top on the right shoulder, or treating a brief poke through the neckline as a break. Waiting for a convincing break reduces that risk, at the cost of acting later. Neither choice removes it.

Textbook vs live charts

Real head and shoulders patterns are rarely as tidy as the diagrams. ChartSchool notes that symmetry is preferred but not an absolute requirement: the right shoulder can be higher, lower, wider or narrower than the left. Necklines slope, peaks are ragged, and the pattern is often only obvious once the neckline has already broken.

In TradingLoot's view, two questions help. Is there a clear prior trend for the pattern to reverse? And can you draw the neckline before the break, rather than fitting it afterwards? Practising on historical data with replay is one way to test whether you would have seen the pattern in time.

Common misreadings

These are ways the pattern is often misread, as TradingLoot sees it.

"Three peaks make a head and shoulders."

Not without a prior uptrend and a confirmed break of the neckline.

"The right shoulder is the signal."

The right shoulder completes the shape. The neckline break is what confirms the pattern.

"The measured move tells you where price will go."

It's a convention for a rough reference point, not a forecast of distance or timing.

"Price will always come back to retest the neckline."

Sometimes it does, often it doesn't. A retest is a possibility, not a step in the pattern.

"A failed pattern means patterns don't work."

Failure is part of how any pattern behaves, and it's information in its own right.

Key takeaways

• A head and shoulders top needs a prior uptrend; the inverse needs a prior downtrend.
• The neckline, through the lows between the peaks, is the line that matters.
• Until the neckline breaks convincingly, the shape is not a completed pattern.
• Volume is supporting evidence, and ChartSchool weights it more heavily in the inverse pattern.
• Retests and measured moves are possibilities and conventions, not certainties.
• A confirmed pattern can fail; a busted pattern is Bulkowski's narrower definition of one kind of failure.

Books and further reading

• Encyclopedia of Chart Patterns by Thomas Bulkowski: how he measured patterns, including how they fail.
• Technical Analysis of the Financial Markets by John J. Murphy: a broad foundation in which price patterns are one of many tools.
• Chart Patterns: the method that applies to every pattern.